structural adaptations for ram ventilation: gill fusions in … · 2012. 12. 31. · structural...

Structural Adaptations for Ram Ventilation: Gill Fusionsin Scombrids and Billfishes

Nicholas C. Wegner,1,2* Chugey A. Sepulveda,3 Scott A. Aalbers,3 and Jeffrey B. Graham1

1Center for Marine Biotechnology and Biomedicine, Marine Biology Research Division, Scripps Institution ofOceanography, University of California San Diego, La Jolla, California 920932Fisheries Resource Division, Southwest Fisheries Science Center, National Marine Fisheries Service, NationalOceanic and Atmospheric Administration, La Jolla, California 920373Pfleger Institute of Environmental Research, Oceanside, California 92054

ABSTRACT For ram-gill ventilators such as tunas andmackerels (family Scombridae) and billfishes (familiesIstiophoridae, Xiphiidae), fusions binding the gill lamellaeand filaments prevent gill deformation by a fast and con-tinuous ventilatory stream. This study examines the gillsfrom 28 scombrid and seven billfish species in order todetermine how factors such as body size, swimmingspeed, and the degree of dependence upon ram ventilationinfluence the site of occurrence and type of fusions. In thefamily Scombridae there is a progressive increase in thereliance on ram ventilation that correlates with the elabo-ration of gill fusions. This ranges from mackerels (tribeScombrini), which only utilize ram ventilation at fastcruising speeds and lack gill fusions, to tunas (tribe Thun-nini) of the genus Thunnus, which are obligate ram venti-lators and have two distinct fusion types (one binding thegill lamellae and a second connecting the gill filaments).The billfishes appear to have independently evolved gillfusions that rival those of tunas in terms of structuralcomplexity. Examination of a wide range of body sizes forsome scombrids and billfishes shows that gill fusionsbegin to develop at lengths as small as 2.0 cm fork length.In addition to securing the spatial configuration of the gillsieve, gill fusions also appear to increase branchial resist-ance to slow the high-speed current produced by ram ven-tilation to distribute flow evenly and optimally to the re-spiratory exchange surfaces. J. Morphol. 000:000–000,2012. � 2012 Wiley Periodicals, Inc.

KEY WORDS: tuna; mackerel; marlin; swordfish; gillfilament; gill lamellae

INTRODUCTION

Tunas, bonitos, and mackerels (familyScombridae) and billfishes (families Istiophoridae,Xiphiidae) are continuous swimmers and breatheusing ram ventilation, the mechanism throughwhich forward swimming provides the forcerequired to drive water into the mouth andthrough the branchial chamber (Brown and Muir,1970; Stevens, 1972; Roberts, 1975, 1978; Stevensand Lightfoot, 1986; Wegner et al., 2012). Ramventilation transfers the energetic cost of activegill ventilation to the swimming musculature, andbecause mouth and opercular motions are mini-

mized, both respiratory and swimming efficiencyare increased (Freadman, 1979, 1981; Steffensen,1985). However, at the relatively high swimmingspeeds attained by scombrids and billfishes, ramventilation poses two challenges. First, the ram-ventilatory current must be slowed to create opti-mal flow conditions for efficient gas exchange atthe respiratory lamellae (Brown and Muir, 1970;Stevens and Lightfoot, 1986; Wegner et al., 2012).Second, the gills must be reinforced in order tomaintain normal orientation with respect to ahigh-pressure branchial stream, the force of whichincreases with swimming speed (Muir and Ken-dall, 1968; Brown and Muir, 1970). Recent work byWegner et al. (2010) showed that the morphomet-rics of scombrid and billfish gills (e.g., shape, size,and number of gill lamellae) increase branchial re-sistance and help slow ram-ventilatory flow. Toincrease the overall rigidity of the branchial sieve,some scombrids and billfishes have structural sup-ports in the form of gill fusions (Muir and Kendall,1968; Muir, 1969; Johnson, 1986; Wegner et al.,2006).

In most teleost fishes, the gill filaments are notinterconnected and extend independently from thegill arch, with lamellae extending freely from the

Contract grant sponsor: National Science Foundation; Contractgrant number: IOS-0817774; Contract grant sponsors: The TunaIndustry Endowment Fund at Scripps Institution of Oceanography,the Pfleger Institute of Environmental Research, the George T.Pfleger Foundation, the Moore Family Foundation, the Nadine A.and Edward M. Carson Scholarship awarded by the AchievementRewards for College Scientists (ARCS), Los Angeles Chapter(N.C.W.), a National Research Council Associateship (N.C.W.), andthe Kennel-Haymet Student Lecture Award (N.C.W.).

*Correspondence to: Nicholas C. Wegner, National Marine Fish-eries Service, Southwest Fisheries Science Center, 8901 La JollaShores Dr., La Jolla, CA 92037. E-mail: [email protected]

Received 4 April 2012; Revised 21 July 2012;Accepted 23 August 2012

Published online inWiley Online Library (wileyonlinelibrary.com)DOI: 10.1002/jmor.20082

JOURNAL OF MORPHOLOGY 000:000–000 (2012)

� 2012 WILEY PERIODICALS, INC.

gill filaments (Fig. 1A,B). The gill fusions of scom-brids and billfishes connect these normally inde-pendent structures to form lamellar fusions (Fig.1C–E) and filament fusions (Fig. 1E). Lamellarfusions, which bind the gill lamellae near theirleading (water-entry) edge and thus secure thespatial integrity of the interlamellar channels(5lamellar pores), can be further categorized intotwo forms. Interlamellar fusions (Fig. 1C) bind ad-jacent lamellae on the same filament and havebeen described for wahoo, Acanthocybium solandri(a non-tuna scombrid), and striped marlin, Kajikiaaudax (Wegner et al., 2006). Complete lamellarfusions (Fig. 1D) connect the lamellae on one fila-ment to the closely positioned and opposing lamel-lae of the adjacent filament, thus providing sup-port to both the gill lamellae and filaments (Muirand Kendall, 1968; Muir, 1969; Wegner et al.,2006). Complete lamellar fusions have beenreported in several tuna species and in the stripedmarlin, where, in some areas of the gills, the inter-lamellar fusions of adjacent filaments join together(i.e., the striped marlin has both complete lamellarand interlamellar fusions; Wegner et al., 2006).

Filament fusions (Fig. 1E) bridge the interfila-ment space in a lattice-like pattern, binding adja-cent filaments on the same gill hemibranch (Muirand Kendall, 1968; Johnson, 1986; Davie, 1990).These fusions, which form on the trailing (water-exit) and often leading (water-entry) edges of thefilaments, have been documented in tunas of thegenus Thunnus, the wahoo, and some billfishes

(Istiophorus, Kajikia, and Xiphias; Muir and Ken-dall, 1968; Johnson, 1986). In the tuna genus,Thunnus, filament fusions are formed by theexpansion of the mucosal filament epithelium(Muir and Kendall, 1968). In contrast, wahoo andbillfish filament fusions are formed by bony epithe-lial toothplates, which on the trailing edges of thefilaments cover cartilaginous connections of the fil-ament rods (Johnson, 1986).

The diversity of gill fusion type and structurecan be expected to reflect interspecific differencesin reliance upon, or specialization for, ram ventila-tion. Although all scombrids and billfishes useram ventilation, basic information about the occur-rence of fusions and their structure is not avail-able for many species, and there are few dataaddressing how fusion structure may change withbody size or the range of swimming speedsemployed by these fishes. Because the familyScombridae demonstrates a progression in adapta-tions for fast, continuous swimming (from leastderived mackerels to most derived tunas) and anassociated increase in reliance on ram ventilation,determination of gill fusion type and pattern alongthis gradient can provide insight into the struc-tural requirements of ram ventilation. This com-parative study thus examines the gills of 28 scom-brid and seven billfish species in order to deter-mine how factors such as body size, swimmingspeed, and the degree of dependence upon ramventilation correlate with the site of occurrenceand elaboration of fusions.

Fig. 1. Basic gill morphology and fusion structure described for tunas and billfishes. (A) First gill arch taken from the left side ofthe branchial chamber of a scombrid showing the gill filaments emanating from the arch. (B–E) Magnified views of black box in Ashowing the leading (water-entry) edges of three adjacent gill filaments (each containing two rows of gill lamellae) with differentfusion types: (B) No gill fusions, (C) Interlamellar fusions, (D) Complete lamellar fusions, and (E) Filament fusions (with completelamellar fusions extending underneath). Water flow direction in B-E is into the page. IF, interlamellar fusion; F, filament; FF filamentfusions; L, lamellae; LF, complete lamellar fusion.

2 N.C. WEGNER ET AL.

Journal of Morphology

MATERIALS AND METHODSGill Tissue Collection and Preservation

Fresh gill tissue samples were collected from 19 scombrid andthree billfish species from various locations around the world byworking with other researchers and fishermen. All collected indi-viduals were sacrificed upon capture by severing the spinal cord atthe base of the skull in accordance with protocol S00080 of the Uni-versity of California, San Diego Institutional Animal Care and UseCommittee. Length measurements [fork length for scombrids,lower jaw fork length for billfishes] were taken, and when possible,each specimen was weighed using a digital scale. In cases wherebody mass could not be directly measured, estimates were madeusing species-specific length-weight regressions (Chatwin, 1959;Wolfe and Webb, 1975; Faruk Kara, 1979; Muthiah, 1985; Ramoset al., 1986; McPherson, 1992; Hsu et al., 2000; Chiang et al., 2004;Beerkircher, 2005; de la Serna et al., 2005; Moazzam et al., 2005;Vieira et al., 2005). Whenever possible, a large size range of indi-viduals from each species was collected in order to examinechanges in fusion structure with body size.

For most specimens, the gills from one or both sides of thebranchial chamber were excised immediately and fixed in 10%formalin buffered in sea water. For large fish, a low-pressureseawater hose was used to keep the gills wet during excision,which, depending on the size of the fish, required up to 10 min(prolonged air exposure results in the degradation of fine gilltissue).

On an opportunistic basis, specimens from select species wereperfused with vascular casting solution in order to determine ifthe elaboration of gill fusions was associated with changes ingill circulation. Vascular casting procedures were performed ontwo yellowfin tuna (Thunnus albacares), one skipjack tuna(Katsuwonus pelamis), seven eastern Pacific bonito (Sarda chi-liensis), three wahoo, one sailfish (Istiophorus platypterus),three striped marlin, and three swordfish (Xiphias gladius).Each freshly killed fish was placed ventral side up in a V-shapedcradle while a low-pressure seawater hose was used to ventilatethe branchial chamber and keep the gills wet. The heart wasexposed by midline incision and cannulated to administer hepari-nized saline (2–5 min) followed by vascular casting solution

TABLE 1. Scombrids and billfishes examined in relation to gill fusion type

Speciesname

Commonname

n total(collected)

Fork length(cm)

Mass(kg)

Lamellarfusion

Filamentfusion

Collectionlocation

Thunnus alalunga Albacore 4 (2) 39.0–82.0 1.0–11.3 LF ME 1Thunnus albacares Yellowfin tuna 25 (19) 29.0–182.0 0.38–94.0 LF ME 1,2,3Thunnus atlanticus Blackfin tuna 11 (5) 45.5–52.0 1.6–2.4 LF ME 4Thunnus obesus Bigeye tuna 6 (5) 40.0–136.0 1.5–46.0 LF ME 3,5Thunnus orientalis Pacific bluefin tuna 5 (5) 86.0–120.0 12.2–32.3 LF ME 2Thunnus tonggol Longtail tuna 9 (2) 38.5–85.5 0.83–9.6 LF ME 6Katsuwonus pelamis Skipjack tuna 14 (8) 30.4–77.0 0.48–9.6 LF -- 1,2,3Euthynnus affinis Kawakawa 14 (4) 19.7–78.0 0.10–6.0 LF -- 6Euthynnus alleteratus Little tunny 2 (0) 6.5–8.0 2.5–4.6 g LF -- —Euthynnus lineatus Black skipjack 22 (10) 10.9–59.0 16.4 g–2.94 LF -- 2,7Auxis rochei Bullet tuna 3 (0) 21.8–36.0 0.11–0.76 LF -- —Auxis thazard Frigate tuna 25 (2) 7.7–44.5 3.4 g–1.48 LF -- 2Allothunnus fallai Slender tuna 2 (0) 84.5–85.5 11.3–11.8 LF -- —Sarda australis Australian bonito 1 (0) 33 0.52 LF -- —Sarda chiliensis Eastern Pacific bonito 25 (25) 27.5–82.0 0.15–6.4 LF -- 1,2Sarda orientalis Striped bonito 2 (0) 48.0–50.5 1.45–1.56 LF -- —Sarda sarda Atlantic bonito 2 (0) 11.5–48.0 10.0 g–1.58 LF -- —Cybiosarda elegans Leaping bonito 7 (7) 30.6–35.5 0.44–0.67 LF -- 6Acanthocybium solandri Wahoo 9 (9) 75.0–152.0 2.1–24.2 IF ETP 2,3Scomberomorus commerson Narrow-barred

Spanish mackerel6 (5) 43.5–96.0 0.59–7.0 -- -- 6

Scomberomorus concolor Monterey Spanishmackerel

5 (0) 38.5–55.0 0.35–1.39 -- -- —

Scomberomorusqueenslandicus

Queenslandschool mackerel

6 (5) 22.3–58.5 65.0 g–1.65 IF* -- 6

Scomberomorus sierra Pacific sierra 5 (5) 47.0–50.0 0.75–0.90 -- -- 7Grammatorcynus bicarinatus Shark mackerel 2 (2) — 3.5–4.5 -- -- 6Grammatorcynus bilineatus Double-lined mackerel 4 (0) 41.5–51.0 0.59–0.95 -- -- —Rastrelliger brachysoma Short mackerel 2 (0) 10.0–10.3 7.5–9.5 g -- -- —Rastrelliger kanagurta Indian mackerel 4 (2) 12.1–31.5 16.6 g–0.43 -- -- 8Scomber japonicus Pacific chub mackerel 25 (25) 9.5–40.0 7.2 g–0.74 -- -- 1Istiophorus platypterus Sailfish 2 (2) 200.0–207.0 31.0–34.4 IF, LF ETP 7Makaira nigricans Blue marlin 3 (0) — 53.2–189.5 ETP —Istiompax indica Black marlin 2 (0) 86.5–90.5 — ETP —Kajikia audax Striped marlin 9 (8) 95.0–188.5 8.0–70.0 IF, LF ETP 1,2Tetrapturus angustirostris Shortbill spearfish 3 (0) 125.0–143.0 — IF, LF ETP —Tetrapturus georgii Roundscale spearfish 1 (0) 173.0 — ETP —Xiphias gladius Swordfish 16 (9) 34.0–223.5 73.1 g–155.3 -- ETP 1

The total number of specimens examined is given with the number collected (i.e., those not from scientific collections) in parenthe-ses. Collection locations: 1. Southern California, USA; 2. Baja California and Baja California Sur, Mexico; 3. Hawaii, USA; 4. GrandCayman Island; 5. Central equatorial Pacific; 6. Queensland, Australia; 7. Costa Rica (Pacific coast); 8. Palau. Abbreviations: ETP,epithelial toothplate filament fusion; IF interlamellar fusion; LF, complete lamellar fusion; ME, mucosal epithelium filament fusion.Dashes (--) indicate a lack of fusions; blank spaces indicate that fusion status remains undetermined. Billfish lengths are given aslower jaw fork lengths. *Interlamellar fusions were present in five S. queenlandicus from 46.9–58.5 cm (0.87–1.65 kg) but were notpresent in a 22.3 cm (65.0 g) specimen. The small yellowfin and sailfish specimens examined to understand the timing of fusion de-velopment are not included in this table. Fish masses are given in kg unless otherwise denoted by grams (g).

GILL FUSIONS FOR RAM VENTILATION 3


(Mercox, Mercox II, Ladd Research, Williston, VT). Gills wereperfused at 70–95 mmHg, which is consistent with vascular cast-ing preformed in other scombrid and billfish studies (Olson et al.,2003; Wegner et al., 2006; Wegner et al., 2010). Following poly-merization, gills from one side of the branchial chamber werefixed to determine fusion status, while the gills from the otherside were stored frozen and then macerated in washes of 15–20%KOH to expose the vascular casts. [Note: Both fixed gill tissueand vascular casts from Wegner et al. (2006) and Wegner et al.(2010) are included in the present analysis].

Gill Samples from Scientific Collections

Preserved specimens from scientific collections at Scripps Insti-tution of Oceanography, the Smithsonian National Museum ofNatural History, the Australian Museum, and the AustralianNational Fish Collection were also examined. Most of these fishhad highly degraded gill tissue (likely caused by prolonged air ex-posure or freezing prior to fixation), which precluded accuratedetermination of fusion status. Because degraded gill tissue hasled to discrepancies in the description of fusions in the literature(Muir and Kendall, 1968; Muir, 1969), only scientific collectionspecimens in which fusion status could be clearly determinedwere included in the present analysis. Although most billfish gilltissue had undergone substantial degradation preventing thedetermination of lamellar fusion status, billfish filament fusionsare composed of tough bony epithelial toothplates, which made itpossible to record their presence in some specimens.

Gill Fusion Assessment

Fixed gill tissue from each specimen was examined to determinethe presence or absence of the different fusion types: filamentfusions (either composed of a mucosal epithelium or bony epithelialtoothplates), complete lamellar fusions, and interlamellar fusions.For most specimens, filament fusion type and the presence of com-plete lamellar fusions could be determined by direct observationwith the naked eye or aided by a dissection microscope. In speci-mens for which complete lamellar fusions were not obvious (e.g.,species with interlamellar fusions or lacking lamellar fusions),scanning electron microscopy (SEM) was used to assess lamellarfusion status. SEM followed the protocol of Wegner et al. (2006).Small sections of gill tissue (usually 1 cm2 or less) were removedfrom each specimen, rinsed in deionized water, and either dehy-drated in 100% ethanol (20–25% increments over 24 h) and criti-cal-point dried, or dehydrated in 100% tert-butyl alcohol (25%increments over 24 h, rinsed twice at 100%), frozen in the alcoholat 48C, and freeze dried. Dried material was then sputter-coatedwith gold-palladium and viewed using an FEI Quanta 600 SEM(FEI, Hillsboro, OR) under high-vacuum mode. Vascular casts ofthe gill circulation were examined using dissection light micros-copy and the SEM under low-vacuummode.

Gill Fusion Development

In order to determine the timing and structural changes asso-ciated with lamellar fusion development, a size range of small

Fig. 2. Scombrid and billfish cladograms showing the presence of the different gill fusion types for each genus. The number ofspecies examined as a ratio of the total number in the genus is given in parentheses. Abbreviations: ETP, epithelial toothplate fila-ment fusion; IF, interlamellar fusions; LF, complete lamellar fusions; ME, mucosal epithelium filament fusion. Dashes (--) indicatea lack of fusions; blank spaces indicate that fusion status remains undetermined. Scombrid cladogram based on Collette et al.(2001); billfish cladogram based on Collette et al. (2006). Included in the number of Thunnus and Scomber species examined is theAtlantic bluefin tuna, T. thynnus and the Atlantic mackerel, S. scombrus, for which fusion data were reported by Muir and Kendall(1968). Lamellar fusion status in Makaira and Istiompax remains undetermined due to the poor preservation of lamellar structurein the scientific collection specimens examined.



individuals from one scombrid species (yellowfin tuna, n 5 6,1.5–5.2 cm, 0.10–2.22 g) and one billfish species (sailfish, n 515, 12.4–33.0 cm, 4.5–95.5 g) were examined.

Statistics

It is important to note that correlations between fusion typeand structure and dependence on ram ventilation are based ongeneral observations only and were not tested statistically. Sta-tistical techniques such as phylogenetic independent contrastcan not be employed as the degree of dependence on ram venti-lation (as determined through measures of respiratory move-ments at varying swimming speeds) has not been quantified formost species examined.

RESULTS

Table 1 shows the presence and type of gill fusions(along with collection location data) for the 28 scom-brid and seven billfish species examined in thisstudy. Fusion status for each genus is mapped ontothe scombrid and billfish phylogenies in Figure 2.

Lamellar Fusions

Scanning electron micrographs depicting thestructural details for complete lamellar and inter-lamellar fusions are shown in Figure 3. Completelamellar fusions (Fig. 3A) are present in all fivetuna genera (i.e., Thunnus, Katsuwonus, Euthyn-nus, Auxis, and Allothunnus) and the bonito generaSarda and Cybiosarda (quality specimens of themonotypic Gymnosarda and Orcynopsis were notavailable for examination). This study further veri-fies interlamellar fusions (Fig. 3B) in striped marlinand wahoo as described previously (Wegner et al.,2006) and also documents their presence in sailfish,shortbill spearfish, Tetrapturus angustirostris, andone Spanish mackerel species, the Queenslandschool mackerel, Scomberomorus queenlandicus(Fig. 3C,D). The interlamellar fusions of Queens-land school mackerel and wahoo are thinner andless complete (i.e., fusions do not bind all of thelamellae along the length of the filament) than

Fig. 3. Scanning electron micrographs of complete lamellar and interlamellar fusions from (A) a 1.45 kg eastern Pacificbonito, (B) a 45.0 kg striped marlin, (C) a 1.07 kg Queensland school mackerel, (D) a 1.07 kg Queensland school mackerel(magnified image of box in C), (E) a 25.0 kg striped marlin, and (F) a 1.45 kg eastern Pacific bonito. B and E are fromWegner et al. (2006). F, filament; IF, interlamellar fusion; L, lamellae; LF, complete lamellar fusion. Water flow direction isinto the page.



those of striped marlin, shortbill spearfish, and sail-fish (cf., Fig. 3C,D with 3B), which are well devel-oped, and, in many cases, fuse together to form com-plete lamellar fusions (Fig. 3E). Other Spanishmackerels of the genus Scomberomorus and Gram-matorcynus, the mackerels (Rastrelliger, Scomber),and the swordfish, Xiphias gladius, lack both formsof lamellar fusions. Although nearly all the lamellaein tunas and bonitos are bound by complete lamel-lar fusions, remnants of interlamellar fusions fromdevelopment (see below) are found at the filamenttips in some species (Fig. 3F).

Filament Fusions

Filament fusions occur in tunas of the genusThunnus, the wahoo, and all the billfish speciessampled (spanning both billfish families and all sixbillfish genera) (Table 1, Fig. 2). For all species,

the lattice-like filament fusions on the trailing(water-exit) edges of the gill filaments occur alongnearly their entire length (i.e., from the gill archto the filament tips). However, on the leading(water-entry) edges, species-specific patterns in fil-ament fusion distribution are prevalent and theseare shown for several species in Figure 4A–G(once filament fusions are fully developed theirsize and location remains relatively conservedwithin a species). These patterns range from threeThunnus species that lack filament fusionsentirely on the leading edge: albacore, Thunnusalalunga, blackfin tuna, Thunnus atlanticus, andlongtail tuna, Thunnus tonggol (not shown in Fig.4), to the swordfish in which the leading filamentedges are bound by fusions along nearly theirentire length (Fig. 4G). Many striped marlin speci-mens have a thick fusion of the filament tips (Fig.4F).

All filament fusions examined conform to thestructural pattern described by Johnson (1986), inwhich the filament fusions of Thunnus are exten-sions of the filament mucosal epithelium, whilethose of the wahoo and billfishes are formed byepithelial toothplates (Table 1, Fig. 2). Vasculargill casts reveal an additional structural differencein that the more robust toothplate and cartilagi-nous-based fusions on the trailing edge of wahooand billfish filaments often contain blood vesselsthat connect the circulation of adjacent filaments.Three types of vascular junctions occur: 1) connec-tions between afferent filamental arteries on adja-cent filaments (Fig. 5A), 2) afferent lamellar arte-rioles that supply blood from one filament to a lim-ited number of lamellae on the adjacent filament(Fig 5B), and 3) nutrient blood vessels (i.e., non-re-spiratory vessels), which likely serve to supportthe extensive fusion latticework (not pictured).Vascular connections between filaments are notfound in the mucosal epithelium filament fusionsof Thunnus or in lamellar fusions.

Fusion Development

Figures 6–8 show the progressive developmentof lamellar fusions in a size series of juvenile yel-lowfin tuna. At 1.5 cm (103 mg) the gill filamentsand lamellae are fully developed but lack fusions(Fig. 6). By 2.0 cm (154 mg), however, some inter-lamellar fusions have developed near the filamenttips (Fig. 7A–E). Interlamellar fusion formationappears to involve the bending of the leading la-mellar lateral edge toward the filament tip until itcontacts the adjacent lamella (Fig. 7C–E). By 3.0–3.2 cm (453–915 mg) interlamellar fusions start togrow together to form complete lamellar fusions(Fig. 8). However, interlamellar fusions persistnear the filament tip, and lamellae near the baseof the filaments remain free of fusions. By 5.5 cm(2.22 g) most of the interlamellar fusions have

Fig. 4. Filament fusions on the leading edge of the anteriorhemibranch of the third gill arch near the cerato-epibranchialjoint for (A) a 32.3 kg Pacific bluefin tuna, (B) a 72.0 kg yellow-fin tuna, (C) a 46.0 kg bigeye tuna, (D) a 24.2 kg wahoo, (E) a34.4 kg sailfish, (F) a 67.8 kg striped marlin, and (G) a 64.9 kgswordfish. F and G are from Wegner et al. (2010).



grown together to form complete lamellar fusions,which have progressed further towards the base ofthe filaments, leaving fewer non-fused lamellae.

In sailfish the interlamellar fusions also firstform near the filament tip and extend toward thefilament base as body size increases. Figure 9shows the distribution of interlamellar fusions onthe gill filaments of a 28.5 cm (68.0 g) sailfish.Although interlamellar fusions were observed overthe entire size range of small sailfish examined(12.4–33.0 cm lower jaw fork length, 4.5–95.5 g),no complete lamellar fusions were observed.

Because of the limited number of specimensavailable over the required size range, the progres-sion and the exact onset of filament fusion devel-opment could not be assessed in the same detail as

lamellar fusions. Among scombrids, filamentfusions consistently appear to begin developmentat relatively small body sizes, first appearing nearthe gill arch and progressing toward the filamenttips with growth. In yellowfin tuna, filamentfusions already bind approximately 40% of thelength of the trailing filament edges by 29 cm(0.38 kg) and 50–90% of the trailing edges by 32.6cm (0.60 kg). By 43.5 cm (1.6 kg) filament fusionsare completely formed on the trailing edges andpartially formed on the leading edges. Likewise, onthe smallest available specimens of longtail (38.5cm, 0.83 kg) and blackfin tuna, (46.0 cm, 1.6 kg),filament fusions are completely formed on thetrailing edges, while in the smallest bigeye tuna(40.0 cm, 1.5 kg) and wahoo (75.0 cm, 2.1 kg), fila-

Fig. 5. Scanning electron microscope images of vascular casts from a 31.0 kg sailfish showing the afferent gill filamentcirculation. (A) Six adjacent afferent filamental arteries (AFA) connected at several locations along their lengths (*). (B)Magnified view of dotted box in A, showing the afferent lamellar arterioles (ALA). The largest ALA (*) bridges the interfila-ment space providing blood from one filament to the lamellae (L) of the adjacent filament.

Fig. 6. Scanning electron micrographs of the gill arches, filaments, and lamellae from a 1.5cm (103 mg) yellowfin tuna.



ment fusions are also beginning to develop on theleading edges. In billfishes, there appear to beinterspecific differences in the timeline of filamentfusion development. Although not present in a33.0 cm (95.5 g) sailfish, filament fusions are al-ready completely formed on the trailing edges ofthe gill filaments in a 34 cm (73.1 g) swordfish. By56.5 cm (1.3 kg), swordfish filament fusions alsobind approximately 30–50% of the leading edgesclosest to the gill arch. By 88.5 cm (6.9 kg) sword-fish filament fusions are fully formed and cover

both the leading and trailing edges as shown inFigure 4.

DISCUSSION

This study of 28 scombrid and seven billfish spe-cies provides data supporting a direct relationshipbetween branchial fusions and reliance upon ramventilation. Although many fish groups utilize ramventilation while swimming at fast velocities (Rob-erts, 1975; 1978; Wegner and Graham, 2010), gill

Fig. 7. Scanning electron micrographs of the gill filaments and lamellae of a 2.0 cm (154 mg) yellowfin tuna. (A) Fila-ments from the first gill arch. (B) Enlarged image of dotted box in A showing interlamellar fusions (IF) forming near somefilament tips. (C) Filament tips with interlamellar fusions. (D) Magnification of dotted box in C (left). (E) Enlarged image ofdotted box in C (right) showing the curving of a lamella toward the filament tip to fuse with the adjacent lamella.



fusions only appear present in species highly de-pendent upon this form of respiration. In thesefishes, the continuous high-pressure ventilatorystream produced through ram ventilation appearsto require fusions to maintain the spatial and struc-tural integrity of the gills and, in some cases, slowand distribute branchial flow evenly to the respira-tory surfaces to optimize gas transfer efficiency.

Gill Fusions and Ram VentilationScombrids. Morphological and physiological

comparison of the four major scombrid tribes(mackerels, tribe Scombrini; Spanish mackerels,tribe Scomberomorini; bonitos, tribe Sardini;tunas, tribe Thunnini) demonstrates a correlationbetween the progressive development of gradedadaptations related for high-performance swim-ming (including a dependence on ram ventilation)and the elaboration of gill fusions (Fig. 2). Themackerels (Scomber and Rastrelliger, tribe Scom-brini) are the most basal members of the scombridsubfamily Scombrinae (Fig. 2). Although therehave been no studies on the respiratory biome-chanics of Rastrelliger, Roberts (1975) showed thatthe Atlantic mackerel, Scomber scombrus, is notan obligatory ram ventilator. S. scombrus usesactive gill ventilation during slow swimming andswitches to ram ventilation at speeds of 53–75 cms21 (2.7–4.7 body lengths s21). While neitherScomber nor Rastrelliger have gill fusions, studies

with S. japonicus (Wegner et al., 2010) have iden-tified two lamellar features also present in morederived scombrids that likely facilitate ram venti-lation when swimming at fast speeds. First, mack-erel gill lamellae have a long rectangular shapethat reduces their profile (i.e., height) and providesan extended surface for attachment to the gill fila-ment, thus enhancing lamellar rigidity. Second,mackerel lamellae are closely spaced (i.e., mack-erel gills have a high lamellar frequency), which,in addition to augmenting the total respiratorysurface area of the gills to power continuous swim-ming, contributes to gas-transfer efficiency by min-imizing physiological dead space and increasingbranchial resistance to slow the high-speed bran-chial current produced by ram ventilation.

The Spanish mackerels (Grammatorcynus,Scomberomorus, and Acanthocybium, tribe Scom-beromorini) are intermediate between the macker-els and the more derived tunas and bonitos (Col-lette et al., 2001). Morphological features such as awell-developed lateral keel on the caudal peduncledistinguish this group from the mackerels andsuggest an increased capacity for fast, sustainableswimming (Collette and Russo, 1984; Colletteet al., 2001). While there are no data on thisgroup’s range of swimming speeds or its depend-ence upon ram ventilation, the finding of interlam-ellar fusions in Queensland school mackerel,Scomberomorus queenslandicus, but not in otherspecies of this genus or in Grammatorcynus sug-

Fig. 8. Scanning electron micrographs of the gill filaments from a 3.2 cm (915 mg) yellowfin tuna. (A) Interlamellarfusions near the filament tips that grow together to form complete lamellar fusions; no fusions are present near the base ofthe filaments. (B) Magnified image of dashed box in A showing complete lamellar and interlamellar fusions. (C) Gill fila-ments with interlamellar fusions near the tips, but no complete lamellar fusions.



gests varying levels of dependence on, and speciali-zation for, ram ventilation. The wahoo, considereda specialized offshoot of Scomberomorus (Colletteet al., 2001), has interlamellar fusions similar tothose of the Queensland school mackerel as wellas filament fusions composed of bony epithelialtoothplates. The more rigid gill sieve of the wahoosuggests it has a greater dependence on ram venti-lation than other members of the tribe. This issupported by the wahoo’s more oceanic distribution(Collette and Nauen, 1983) in comparison to otherScomberomorini, and that its burst swimmingspeeds are comparable to those of large tunas andbillfishes (Walters and Fierstine, 1964), groupsthat also have filament fusions.

Regarded as sister groups, tunas (tribe Thun-nini) and bonitos (tribe Sardini) are the mostderived scombrids and share several morphologicaland physiological features related to high-perform-ance swimming (Collette et al., 2001; Graham andDickson, 2004). Although tunas have a greaterdegree of physiological and biochemical specializa-tion [e.g., regional endothermy, greater enzymaticactivities, higher metabolic rates, and larger gillsurface areas (Korsmeyer and Dewar, 2001; Sepul-veda et al., 2003; Graham and Dickson, 2004;Wegner et al., 2010)], both tunas and bonitos are

obligate ram ventilators (Brown and Muir, 1970;Sepulveda et al., 2003). All of the tuna and bonitospecies examined in this study possess completelamellar fusions, which by connecting lamellae onadjacent filaments (Fig. 3A), bind lamellar poredimensions. This increases gas-transfer efficiencyby limiting non-respiratory shunting associatedwith anatomical dead space (i.e., the high-pressureram-ventilatory stream cannot push apart the gilllamellae and filaments to bypass the respiratoryexchange surfaces). The addition of filamentfusions in the most derived tuna genus, Thunnus(Fig. 4), further enhances the rigidity provided bylamellar fusions and appears to correlate with thegenerally larger body sizes of species in this genus(and hence their faster cruise speeds) in compari-son to other tunas and bonitos. Within Thunnus,the smaller species (i.e., albacore, blackfin tuna,longtail tuna) only have filament fusions on thetrailing edges of the gill filaments. However, in thelarger body-sized species (e.g., yellowfin, bigeyetuna, Thunnus obesus, and Pacific bluefin tuna,Thunnus orientalis) filament fusions are also pres-ent on the leading edges (Fig. 4A–C).

Billfishes. All billfishes reach relatively largebody sizes and are all thought to be obligate ramventilators. As a result, a progressive trend in the

Fig. 9. Gill filaments from a 28.5 cm (68.0 g) sailfish. (A) Synoptic view of the entire length of the gill filaments emanat-ing from the gill arch on the left. (B) Enlarged image of dotted box in A (left) showing non-fused lamellae near the base ofthe filaments. (C) Magnified box in A (right) showing interlamellar fusions near the filament tips.



complexity of gill fusions is not evident within thisgroup. All of the seven species examined havetoothplate-based filament fusions that are presenton both the leading and trailing edges of the fila-ments. However, some differences occur in lamel-lar fusion status; sailfish, shortbill spearfish, andstriped marlin gills are similar in the presence ofcomplete lamellar and interlamellar fusions, whilethe swordfish lacks both. Lamellar fusion statusremains unresolved in Makaira and Istiompax dueto the poor quality of the lamellar structure of thepreserved specimens examined. However, based onthe billfish phylogeny shown in Figure 2 and thepresence of interlamellar and complete lamellarfusions in striped marlin, sailfish, and shortbillspearfish, these two types of lamellar fusion arelikely also present in Makaira and Istiompax.

Gill Fusion Development and Correlationswith Body Size

The findings of interlamellar fusions in a 2.0 cmyellowfin tuna (Fig. 7) and complete lamellarfusions in a 3.2 cm specimen (Fig. 8) are consistentwith the observation of complete lamellar fusions ina 3 cm skipjack tuna (Muir and Kendall, 1968) andindicate that tunas and bonitos make an early tran-sition to the use of, and possible reliance upon, ramventilation. Also supporting this contention are thefindings of complete lamellar fusions in the smallestspecimens (6–12 cm) of other tuna and bonito spe-cies examined (Table 1). In contrast, the occurrenceof interlamellar rather than complete lamellarfusions in 12.5–33.0 cm sailfish suggests a slowertransition to a complete reliance upon ram ventila-tion in billfishes. The lack of interlamellar fusionsin a small (22.7 cm) Queensland school mackerel,but their presence in larger individuals (46.9–58.5cm), suggests a similar trend (Table 1).

Gill Fusions and Water Flow

The complete development of filament fusions inThunnus, wahoo, and billfishes at relatively smallbody sizes suggests their function in more thanjust gill support. Muir and Kendall (1968) pro-posed that filament fusions may also aid respira-tion by optimizing both the speed and volume ofwater entering the lamellar channels. In this way,filament fusions would act in concert with modifi-cations in lamellar shape and frequency to opti-mize water flow to the gills for gas exchange (i.e.,long and closely spaced lamellae resulting in long,narrow interlamellar channels increases branchialresistance and slows flow; Hughes, 1966; Wegneret al., 2010). The finding of a negative correlationbetween lamellar frequency and the prevalence offilament fusions suggests that the added resistanceof filament fusions may relax selection for narrowinterlamellar channels (Wegner et al., 2010). For

example, lamellar frequency is highest in scom-brids lacking filament fusions (e.g., Pacific chubmackerel, eastern Pacific Bonito, and skipjacktuna) and lowest in the swordfish, in which fila-ment fusions cover both the entire leading andtrailing edges of the gill filaments (Fig. 4G). Otherscombrids and billfishes appear to span the gamutin between: yellowfin tuna and wahoo, which havefilament fusions along 30–40% of their leading fila-ment edges (Fig. 4B,D), have lower lamellar fre-quencies than skipjack, but higher lamellar fre-quencies than the striped marlin (Wegner et al.,2010), which has filament fusions covering 70–80%of its leading filament edges (Fig. 4F).

The role of filament fusions in optimizing venti-latory flow rates to the respiratory lamellae mayhelp to explain their distribution on the leadingedges of the gill filaments. With the exception ofthe swordfish, in which filament fusions bind theentire leading edges (Fig. 4G), fusions in other spe-cies are most abundant near the gill arch (Fig.4A–F). This appears to correlate with the area ofhighest water inertia, and filament fusions at thislocation may help to disperse the flow of waterevenly along the length of the filaments (i.e.,fusions would limit the volume of water enteringthe interlamellar channels near the gill archwhere flow is strongest). The distinctions in fila-ment fusion patterns between species may there-fore reflect interspecific differences in the bran-chial stream. For example, in the smaller bodiedThunnus species (i.e., albacore, blackfin tuna,longtail tuna) water inertia in the branchial cham-ber may be lower than in the larger members ofthis genus swimming at faster speeds, and thus fil-ament fusions may not be required on the leadingedges to help evenly disperse water flow. In addi-tion, the smaller bodied tunas have shorter fila-ments over which water must be dispersed.

Evolution of Gill Fusions

Figure 2 suggests that filament fusions have in-dependently evolved three times for use in ramventilation: once in the billfishes, and twice inscombrids (in the wahoo, and again in the genusThunnus). The similarity of the epithelial tooth-plate-covered filament fusions of the wahoo andbillfishes, along with a number of other sharedcharacters, led Johnson (1986) to propose that thebillfishes are sister group to the wahoo and shouldbe included within the family Scombridae. Thiswould indicate the independent evolution of fila-ment fusions occurred only twice. However, recentmolecular work (Orrell et al., 2006) suggests sepa-rate billfish and scombrid suborders and thus sup-ports the independent evolution of these structuresthree times.

The number of evolutionary appearances of la-mellar fusions for ram ventilation remains less



clear. One possibility is that lamellar fusions haveindependently evolved twice, once in scombridsand once in billfishes. Under this scenario, theinterlamellar fusions of Scomberomorus and thewahoo would be the more basal character statethat led to the complete lamellar fusions of bonitosand tunas. This hypothesis is supported by the on-togeny of yellowfin lamellar fusion development inthat lamellae are first bound by interlamellarfusions which subsequently grow together to formcomplete lamellar fusions. However, this hypothe-sis implies the loss of interlamellar fusions inGrammatorcynus and some members of Scombero-morus. While scombrid systematics are fairly wellunderstood (Collette et al., 2001), the additionalcharacter states of complete lamellar and inter-lamellar fusions could be considered in future phy-logenetic analyses.

Conclusions for Scombrid and BillfishSpecializations for Ram Ventilation

Gill adaptations for ram ventilation mustincrease structural rigidity to maintain gill config-uration and moderate the high-pressure ventila-tory stream to create optimal flow conditions inthe interlamellar channels for gas exchange. Theresults of this study combined with gill morpho-metric data (Wegner et al., 2010) show that theprogressive changes in the gill structure of bothscombrids and billfishes for ram ventilationinclude 1) lamellae that are long and low in profile(height) which creates an extended axis for attach-ment to the gill filament and increases lamellarstability, 2) a high lamellar frequency which worksin conjunction with long lamellae to increase bran-chial resistance and slow ram-ventilatory flow soas to optimize gas exchange, 3) binding of adjacentlamellae on the same filament (interlamellarfusions) to secure interlamellar spacing, 4) bindingof interlamellar fusions to form complete lamellarfusions which further strengthens lamellar poreintegrity and increases filament rigidity, and 5)formation of filament fusions that provide addi-tional support to the gills, help to distribute flowevenly along the gill filaments, and increase bran-chial resistance to slow ram-ventilatory flow, con-sequently relieving selection for a high lamellarfrequency.

ACKNOWLEDGMENTS

This work would not have been possible withoutthe many scientists and fishermen that helped inthe acquisition of gill samples; The authors thankN. Ben-Aderet, D. Bernal, R. Brill, D. Cartamil, P.Davie, K. Dickson, J. Finan, D. Fuller, A. Graham,S. Griffiths, R.K. Kopf, D. McCarthy, M.McGrouther, M. Musyl, J. Pepperell, K. Schaefer,H.J. Walker, L. Williams, J. Young, and the crews of

the Polaris Supreme and Oscar Elton Sette. Juve-nile yellowfin tuna samples were donated by JeanneWexler from the Inter-American Tropical TunaCommission’s Achotines Laboratory. Pacific bluefintuna gills were provided compliments of Eric Peder-sen. They additionally thank E. York of the SIO An-alytical Facility for technical assistance with mi-croscopy work and H. Dewar, P. Hastings, M.McHenry, F. Powell, C. Purcell, R. Rosenblatt, andthree anonymous reviewers for their constructivefeedback on this manuscript.

LITERATURE CITED

Beerkircher LR. 2005. Length to weight conversions for wahoo,Acanthocybium solandri, in the Northwest Atlantic. Col VolSci Pap ICCAT 58:1616–1619.

Brown CE, Muir BS. 1970. Analysis of ram ventilation of fishgills with application to skipjack tuna (Katsuwonus pelamis).J Fish Res Bd Can 27:1637–1652.

Chatwin BM. 1959. The relationships between length andweight of yellowfin tuna (Neothunnus macropterus) and skip-jack tuna (Katsuwonus pelamis) from the Eastern TropicalPacific Ocean. Inter-Am Trop Tuna Comm Bull 3:305–352.

Chiang W-C, Sun C-L, Yeh S-Z, Su W-C. 2004. Age growth ofsailfish (Istiophorus platypterus) in waters off eastern Tai-wan. Fish Bull 102:251–263.

Collette BB, McDowell JR, Graves JE. 2006. Phylogeny ofRecent billfishes (Xiphioidei). Bull Mar Sci 79:455–468.

Collette BB, Nauen CE. 1983. FAO species catalogue, Vol 2.Scombrids of the world: An annotated and illustrated cata-logue of tunas, mackerels, bonitos, and related species knownto date. FAO 125:1–137.

Collette BB, Reeb C, Block BA. 2001. Systematics of the tunasand mackerels (Scombridae). In: Block BA, Stevens ED, edi-tors. Tuna: Physiology, Ecology and Evolution. San Diego:Academic Press. pp 5–30.

Collette BB, Russo JL. 1984. Morphology, systematics, and biol-ogy of the Spanish mackerels (Scomberomorus, Scombridae).Fish Bull 82:545–692.

Davie PS. 1990. Pacific Marlins: Anatomy and Physiology. Pal-merston North, New Zealand: Simon Print.88 p.

de la Serna JM, Ortiz de Urbina JM, Alot E, Garcıa S, Rioja P.2005. Biological parameters of bullet tuna (Auxis rochei)observed in the Spanish Mediterranean fisheries. Col Vol SciPap ICCAT 58:517–526.

Faruk Kara O. 1979. Observations on growth and relationshipbetween length and weight of Sarda sarda (Bloch). Inv Pesq43:95–105.

Freadman MA. 1979. Swimming energetics of striped bass(Morone saxatilis) and bluefish (Pomatomus saltatrix): Gillventilation and swimming metabolism. J Exp Biol 83:217–230.

Freadman MA. 1981. Swimming energetics of striped bass(Morone saxatilis) and bluefish (Pomatomus saltatrix): Hydro-dynamic correlates of locomotion and gill ventilation. J ExpBiol 90:253–265.

Graham JB, Dickson KA. 2004. Tuna comparative physiology. JExp Biol 207:4015–4024.

Hsu CC, Liu HC, Wu CL, Huang ST, Liao HK. 2000. New infor-mation on age composition and length-weight relationship ofbluefin tuna, Thunnus thynnus, in the southwestern NorthPacific. Fish Sci 66:485–493.

Hughes GM. 1966. The dimensions of fish gills in relation totheir function. J Exp Biol 45:177–195.

Johnson GD. 1986. Scombroid phylogeny: An alternative hy-pothesis. Bull Mar Sci 39:1–41.

Korsmeyer KE, Dewar H. 2001. Tuna metabolism and ener-getics. In: Block BA, Stevens ED, editors. Tuna: Physiology,Ecology and Evolution. San Diego: Academic Press. pp 35–78.



McPherson GR. 1992. Age and growth of the narrow-barredSpanish mackerel (Scomberomorus commerson Lacepede,1800) in north-eastern Queensland waters. Aust J Mar Fresh-water Res 43:1269–1282.

Moazzam M, Osmany HB, Zohra K. 2005. Indian mackerel(Rastrelliger kanagurta) from Pakistan-I. Some aspects ofbiology and fisheries. Rec Zool Surv Pakistan 16:58–75.

Muir BS. 1969. Further observations on gill modifications ofoceanic fishes. Copeia 1969:629.

Muir BS, Kendall JI. 1968. Structural modifications in the gillsof tunas and some other oceanic fishes. Copeia 1968:388–398.

Muthiah C. 1985. Fishery and bionomics of tunas at Mangalore.CMFRI Bull 36:51–70.

Olson KR, Dewar H, Graham JB, Brill RW. 2003. Vascularanatomy of the gills in a high energy demand teleost, theskipjack tuna (Katsuwonus pelamis). J Exp Zool 297A:17–31.

Orrell TM, Collette BB, Johnson GD. 2006. Molecular data sup-port separate scombroid and xiphioid clades. Bull Mar Sci79:505–519.

Ramos A, Alot E, Caminas JA. 1986. Relacion talla/peso de lamelva, Auxis thazard, para el Atlantico y Mediterraneo. ColVol Sci Pap ICCAT 25:265–268.

Roberts JL. 1975. Active branchial and ram gill ventilation infishes. Biol Bull 148:85–105.

Roberts JL. 1978. Ram gill ventilation in fish. In: Sharp GD,Dizon AE, editors. The Physiological Ecology of Tunas. SanFrancisco: Academic Press. pp 83–88.

Sepulveda CA, Dickson KA, Graham JB. 2003. Swimming per-formance studies on the eastern Pacific bonito Sarda chilien-sis, a close relative of the tunas (family Scombridae) I. Ener-getics. J Exp Biol 206:2739–2748.

Steffensen JF. 1985. The transition between branchial pumpingand ram ventilation in fishes: Energetic consequences and de-pendence on water oxygen tension. J Exp Biol 114:141–150.

Stevens ED. 1972. Some aspects of gas exchange in tuna. J ExpBiol 56:809–823.

Stevens ED, Lightfoot EN. 1986. Hydrodynamics of water flowin front of and through the gills of skipjack tuna. Comp Bio-chem Physiol 83A:255–259.

Vieira KR, Oliveira JEL, Barbalho MC, Aldatz JP. 2005.Aspects of the dynamic population of blackfin tuna (Thunnusatlanticus - Lesson, 1831) caught in the northeast Brazil. ColVol Sci Pap ICCAT 58:1623–1628.

Walters V, Fierstine HL. 1964. Measurements of swimmingspeed of yellowfin tuna and wahoo. Nature 202:208–209.

Wegner NC, Graham JB. 2010. George Hughes and the historyof fish ventilation: From Du Verney to the present. Comp Bio-chem Physiol 157A:1–6.

Wegner NC, Lai NC, Bull KB, Graham JB. 2012. Oxygen utili-zation and the branchial pressure gradient during ram venti-lation in the shortfin mako, Isurus oxyrinchus: Is lamnidshark-tuna convergence constrained by elasmobranch gillmorphology? J Exp Biol 215:22–28.

Wegner NC, Sepulveda CA, Bull KB, Graham JB. 2010. Gillmorphometrics in relation to gas transfer and ram ventilationin high-energy demand teleosts: Scombrids and billfishes. JMorphol 271:36–49.

Wegner NC, Sepulveda CA, Graham JB. 2006. Gill specializa-tions in high-performance pelagic teleosts, with reference tostriped marlin (Tetrapturus audax) and wahoo (Acantho-cybium solandri). Bull Mar Sci 79:747–759.

Wolfe DC, Webb BF. 1975. Slender tuna (Allothunnus fallai Ser-venty): First record of bulk catches, Tasmania 1974. Aust JMar Freshwater Res 26:213–221.



structural adaptations for ram ventilation: gill fusions in … · 2012. 12. 31. · structural...

Documents